Deep blank-field catalogue for medium- and large-sized telescopes

Transcription

1 Mon. Not. R. Astron. Soc. 427, (2012) doi: /j x Deep blank-field catalogue for medium- and large-sized telescopes F. M. Jiménez-Esteban, 1,2,3 A. Cabrera-Lavers, 4,5 N. Cardiel 6 and J. M. Alacid 1,2 1 Centro de Astrobiología (INTA-CSIC), Departamento de Astrofísica, PO Box 78, E-28691, Villanueva de la Cañada, Madrid, Spain 2 Spanish Virtual Observatory, Spain 3 Saint Louis University, Madrid Campus, Division of Science and Engineering, Avenida del Valle 34, E Madrid, Spain 4 Instituto de Astrofísica de Canarias, E La Laguna, Tenerife, Spain 5 GTC Project, E La Laguna, Tenerife, Spain 6 Departamento de Astrofísica y CC. de la Atmósfera, Facultad de Ciencias Físicas, Avenida Complutense s/n, E Madrid, Spain Accepted 2012 August 7. Received 2012 August 2; in original form 2012 May 16 1 INTRODUCTION In observational astrophysics, and in particular focusing on imaging acquisition through the optical window, data treatment is fundamental to minimize the influence of data acquisition imperfections on the estimation of the desired astronomical measurements (e.g. Gilliland 1992). For this purpose, image flat-fielding and sky subtraction constitute two of the most common and important reduction steps. An inadequate flat-fielding or sky subtraction easily leads to the introduction of systematic uncertainties in the data. Therefore, the observation of blank fields (BFs), defined as regions of the sky that are devoid of stars down to a given threshold magnitude (m th ), is a very important aspect in astronomical observations. In our previous paper (Cardiel et al. 2011, hereafter Paper I), we have already discussed in detail the relevance of BFs in astronomical observations, and we refer the interested reader to the introduction of that paper. In Paper I, we presented the first systematic all-sky catalogue of BFs so far. The method used to identify the BFs was based on Delaunay triangulation on the surface of a sphere, using stars brighter than 11 mag taken from the Tycho-2 (Høg et al. 2000) catalogue as the nodes for triangulation. In addition, a new virtual observatory fran.jimenez-esteban@cab.inta-csic.es ABSTRACT The observation of blank fields, defined as regions of the sky that are devoid of stars down to a given threshold magnitude, constitutes one of the most relevant calibration procedures required for the proper reduction of astronomical data obtained following typical observing strategies. In this work, we have used Delaunay triangulation to search for deep blank fields throughout the whole sky, with a minimum size of 10 arcmin in diameter and an increasing threshold magnitude from 15 to 18 in the R band of the USNO-B Catalog of the United States Naval Observatory. The result is a catalogue with the deepest blank fields known so far. A short sample of these regions has been tested with the 10.4-m Gran Telescopio Canarias, and it has been shown to be extremely useful for medium- and large-sized telescopes. Because some of the regions found could also be suitable for new extragalactic studies, we have estimated the galactic extinction in the direction of each deep blank field. This catalogue is accessible through the virtual observatory tool TESELA, and the user can retrieve and visualize using ALADIN the deep blank fields available near a given position in the sky. Key words: methods: data analysis methods: numerical catalogues virtual observatory tools. (VO) tool called TESELA, 1 accessible through the internet, was created so that users can retrieve, and visualize, the BFs available near a given position in the sky. Another commonly used resource is the collection by Marco Azzaro, 2 consisting of a short list of 38 BFs void of stars up to mag. Although, in general, deeper than the BFs of Paper I, these are very few, and most of them are in the Northern hemisphere; there are only a few with negative declination, and none below 20.So, this BF collection is not well suited for telescopes located in the Southern hemisphere. Nevertheless, even the Azzaro BFs have been shown to be shallow for large-class telescopes. For example, in the case of the Gran Telescopio Canarias 3 (GTC), a 10-m class telescope, equipped with the Optical System for Imaging and Low Resolution Integrated Spectroscopy (OSIRIS; 4 Cepa et al. 2000), it is usually necessary to have access to BFs free of stars down to 17.5 mag. This requirement C 2012 The Authors

2 680 F. M. Jiménez-Esteban et al. will be even more demanding for the future generation of telescopes, such as the European Extremely Large Telescope 5 (E-ELT). One method of identifying low-populated fields is to use a binnedup smoothed map of the distribution of the stars in the sky. In this method, it is necessary to divide the sky into small patches and to identify those with few or even no stars, down to a threshold magnitude. However, with this method, a BF would not be identified unless the patch was well centred on the region free of stars. In addition, the size of the BF would not be the maximum possible because it is arbitrarily defined. Thus, the procedure should be repeated many times with both a different centring scheme and a different patch radius; this would be very demanding with respect to CPU time. These two drawbacks can be overcome with the sky tessellation method we have developed, which has been extensively described in Paper I. Thus, in this work we have applied Delaunay triangulation to the USNO-B Catalog of the United States Naval Observatory, and we have generated a catalogue of deep blank fields (DBFs) to varying limiting magnitudes from 15 to 18 in steps of 1 mag. In order to validate our method, we have tested some DBFs found at the GTC and we have compared these with the BFs in the Azzaro catalogue. Finally, in order to facilitate the use of this new catalogue, we have incorporated it into TESELA, which now provides a simple interface that allows the user to retrieve the list of either the BFs from Tycho-2 or the DBFs from the USNO-B Catalog, available near a given position in the sky. In Section 2, we describe the method used to create the catalogue, the defects found in the USNO-B Catalog and the results of the tessellation. In Section 3, we describe the use of the DBF catalogue on the GTC. Finally, in Section 4, we describe the implementation of TESELA. The catalogue will be available to download from CDS at 2 THE CATALOGUE 2.1 Tessellating the sky Delaunay triangulation (Delaunay 1934) consists of the subdivision of a geometric object (e.g. a surface or a volume) into a set of simplices. In particular, for the Euclidean planar (two-dimensional) case, given a set of points, also called nodes, Delaunay triangulation becomes a subdivision of the plane into triangles, whose vertices are nodes. For each of these triangles, it is possible to determine its associated circumcircle, which is the circle passing exactly through the three vertices of the triangle. Interestingly, in a Delaunay triangulation, all the triangles satisfy the empty circumcircle interior property, which means that all the circumcircles are empty (i.e. there are no nodes inside any of the computed circumcircles). See Paper I for an extended description of the method. Delaunay triangulation can be applied to the two-dimensional surface of a three-dimensional celestial sphere, using as nodes the locations of the stars down to a given m th. Then, the empty circumcircle interior property of the Delaunay triangulation provides a straightforward method for a systematic search of regions void of stars brighter than a certain magnitude. In this work, we have applied Delaunay triangulation to the sources of the USNO-B Catalog (Monet et al. 2003). The USNO-B Catalog contains astrometric and photometric information for more than a billion stars and galaxies in the sky. The data were taken from various sky surveys over 50 years, and it is expected to be complete up to magnitude V = 21 mag. Typical uncertainties are 200 mas for the positions at J2000 and 0.3 mag in photometry. Photometric data consists of V, R and I optical passbands. We have used the R band (between 600 and 750 nm) as the magnitude reference. For most of the sources, the USNO-B Catalog provides the R magnitude (m R ) at two different epochs. To assign a single m R to these objects, either we have averaged out both m R when their values are similar ( m R < 1) or we have selected the lower value (i.e. the brightest value) when there is a larger difference. In principle, it is straightforward to obtain a list of BFs for the whole celestial sphere by using as input a stellar catalogue including stars down to a given magnitude. However, as explained in Paper I, because of the exponential growth of the number of objects with increasing limiting magnitude, tessellating the whole sky with m th above 11 mag requires too many computer resources (in the form of memory and computing time), even with the approach of subdividing the sky into many smaller subregions. Therefore, to create the DBF catalogue, we have followed a methodology that is different from the one adopted in Paper I. Instead of tessellating the whole sky at once, we used the following workflow. (i) We selected all BFs with a size larger than 10 arcmin, using the deepest catalogue available at that time. (ii) We individually tessellated each of these regions, as in Paper I, using the positions of the USNO-B objects down to a m th as the nodes of the triangulation. (iii) To create the new DBF catalogue, from the new collection of deeper BFs we selected those with a size larger than 10 arcmin. (iv) We repeated the whole process, increasing m th from15to18instepsof1mag. The initial collection of BFs was obtained from Paper I, consisting of BFs void of stars down to 11 mag in both Tycho-2 passbands. This provided initial regions with a size larger than 10 arcmin. This size was selected to ensure that the regions were always larger than the field of view of the OSIRIS instrument used at the GTC (see Section 3). Tessellating this huge number of initial regions requires months of execution time. In order to obtain the results as soon as possible so they could be used in the nightly operation of the GTC, we proceeded with the Northern (Dec. > 30 )skyinstrips1h wide in RA. When the tessellation of the Northern sky was finished, we tessellated the Southern (Dec. 30 ) sky in the same way. Finally we merged all DBFs found into one catalogue. It was very common that the tessellation process identified two overlapping DBFs whose centres were separated by less than 0.3 times the smaller DBF radius. Although both DBFs are correct, the two refer to almost the same region of the sky. Thus, to facilitate the use of the catalogue, we removed the DBF with the smaller radius from the catalogue, retaining the one with the larger size. A final evaluation of each DBF of the catalogue was made. We counted the total number of stars inside the DBF with both m R lower than 17.5 mag (a requirement of the GTC; see Section 3) and lower than 22.0 mag. Then, we estimated the integrated R flux (F int )as the sum of the individual fluxes of all stars in the DBF with m R < 22.0 mag. These quantities provided a measurement of the depth of the DBF. To verify the validity of the method, and taking advantage of the ALADIN 6 software (Bonnarel et al. 2000), we visually checked all 79 DBFs of 18 m th found previously in order to give them to the operation staff of the GTC. This VO tool allows users to visualize and

3 Deep blank-field catalogue 681 analyse digitized astronomical images, and to superimpose entries from astronomical catalogues or data bases available from the VO services. Using ALADIN, we displayed the DSS-red images of the DBF region and we superimposed the USNO-B data of the region. This exercise was very useful in the detection of unknown defects in the USNO-B Catalog (see Section 2.2). Of the 79 DBFs, five corresponded to a dark cloud, so they were removed from the final catalogue. Similarly, we removed any DBFs that corresponded to the same dark cloud of any m th. 2.2 Defects in the USNO-B Catalog The USNO-B Catalog was constructed from photographic plates, taken from several different surveys, which were uniformly scanned, and whose sources were extracted in an automatic way from the scanned data. The original plate images contained many defects and artefacts that logically affected the final USNO-B data. That said, so far there is no other full sky catalogue deep enough for our project. So, the USNO-B Catalog was our best option, even with its known defects. Because we have used the USNO-B Catalog as the input for the tessellation process, our DBF catalogue was affected by the defects of the USNO-B Catalog. Although we have made an effort to minimize the damage, we cannot completely guarantee the absence of misidentification in our final DBF catalogue. So, the user should confirm the validity of the DBFs before using them. During the first inspection of the sky distribution of the DBFs found, we realized that there were abnormal concentrations of DBFs in two rectangular-shaped regions, one around (α J2000, δ J2000 ) = (30,00 ) and the other around (27, 37 ); see Fig. 1. This high density of DBFs reflects an abnormally high value of m R in the USNO-B Catalog, probably because of a wrong photometric calibration of the photographic plates. In Fig. 2, we show a comparison between USNO-A Rmag and USNO-B R1mag and R2mag of the objects in common in a small region of the sky where an abnormally high density of DBFs was found. The values of USNO-B R1mag and R2mag are usually more than 7 mag higher than those of USNO-A Figure 1. Abnormal concentration of DBFs with m th = 15 in the region of the sky of centred at (α J2000, δ J2000 ) = (30,00 ). Figure 2. Comparison between USNO-A Rmag and USNO-B R1mag (red circles) and R2mag (blue triangles) magnitudes of the common objects in the region of the sky centred at (α J2000, δ J2000 ) = (31. 9, 37. 1) and with a radius of 8.5 arcmin. The dashed line depicts the one-to-one correlation. Table 1. Regions where USNO-A Catalog was used instead of USNO- B Catalog. RA ( ) Dec. ( ) (J2000) min max min max Rmag. Thus, the DBFs found in this region are clearly unreliable. Then, we defined four rectangular regions (see Table 1) covering the areas with an abnormally high density of DBFs and we used the USNO-A Catalog in the tessellation instead of the USNO-B Catalog. Barron et al. (2008) have found that 2.3 per cent of the USNO- B Catalog entries do not correspond to real astronomical sources; they are mainly but not only spurious entries, caused by diffraction spikes and circular reflection haloes around bright stars in the original imaging data (see figs 1 and 2 of their paper). In our search for DBFs, the effect of these spurious entries is clear; some of the DBFs might not have been found during the process because of the presence of one or more spurious sources in the field. Thus, these spurious entries might have decreased the number of DBFs, but they have never introduced unreal DBF detections. However, there are other defects in the USNO-B Catalog that could affect our tessellation process in a worse way. In Fig. 3, we show the USNO-B Catalog entries around Vega, which is one of the brightest stars in our sky. The brightest stars (<7 mag)are surrounded by a circular halo in the plate images, caused by internal reflections in the camera. Consequently, the regions around the brightest stars in the USNO-B Catalog are artificial regions that are void of stars, and in many cases these have been wrongly identified as DBFs. With the help of the scripting capabilities of ALADIN, wehave estimated the radius of the region devoid of neighbouring stars around each of the USNO-B stars down to m R = 7.0 mag. For this purpose (see Fig. 4), we consider that around the bright stars

4 682 F. M. Jiménez-Esteban et al. Figure 3. USNO-B Catalog entries superimposed on to the DSS-red arcmin 2 image centred on Vega. where N tot is the total number of neighbouring stars in the annulus between R min and R max. Because we are interested in obtaining an automatic and robust estimation of R min, we have followed an approach based on the incomplete method of Thail (e.g. De Muth 2006). As with most non-parametric procedures, all the objects in the field (within R max ) were first sorted in ascending order using the distance from the central bright star. This provided a collection of N tot points of coordinates (R i, N i ), where i = 1, 2, 3,..., N tot. Here, R i is the distance from the centre and N i is the number of objects within the circle of radius R i. From any pair of such points, and considering equation (1), it is possible to obtain an estimation of R min as Rmin 2 = N bra 2 N arb 2, (2) N b N a where R min R a < R b R max. In this way, it is possible to compute N tot /2 values for R min by using the last equation on the pairs of points where a = 1, 2, 3,..., N tot /2 and b = a + N tot /2 (in fact, this is the basics of the Thail method). Finally, we have determined the median of the N tot /2 estimates of R min. In Fig. 5, we show the variation of the size of the region devoid of neighbouring stars, computed in the way just explained, with m R. On the basis of this diagram, we have applied the following conservative criterion to avoid false DBFs in our catalogue: we have removed all DBFs whose centres were closer than 25 arcmin to stars with m R 0.5, 15 arcmin to stars with 0.5 < m R 3.0 and 10 arcmin to stars with 0.3 < m R 6.0. This criterion was not applied to the DBF with 18 m th, because these were all visually inspected. The percentages of removed DBFs are 1, 2 and 9 per cent for 15, 16 and 17 m th, respectively. As expected, the percentage increases with m th because the number of wrong DBFs should be similar for all catalogues. In Fig. 6, we show another defect of the USNO-B Catalog. There are no entries in the circular region of 9 arcmin diameter around the position in the sky (α J2000, δ J2000 ) = (07:26:40, 78:53:50). Our tessellation process detected this region as void of stars and Figure 4. Illustration of the parameters defined for the automatic determination of the radii around bright stars, which are devoid of neighbouring (fainter) stars. The parameter sought is R min,wherer max is the maximum region around the bright object where the star density is determined. The basic hypothesis of the method is the assumption that the stellar density in the annulus between R min and R max is roughly constant. there is a region devoid of stars up to a radius R min. The automatic determination of this parameter is not difficult if we assume that the stellar density beyond this radius, and up to R max (which, in our case, was set to 30 arcmin), is constant. Under this circumstance, it is easy to show that the number of stars within a radius r (with r > R min )isgivenby r 2 Rmin 2 N(< r) = N tot Rmax 2, (1) R2 min Figure 5. Variation of the size of the artificial void-of-stars regions with the USNO-B R1mag. For stars brighter than m R = 2 mag (small black dots), the radius devoid of neighbouring stars is shown for each individual star. For stars fainter than m R = 2 mag, the median values of the radii are plotted as solid blue circles. The magenta lines indicate the percentiles 15.9 and 84.1 (which approximately correspond to ±1σ confidence limits in a normal distribution).

5 Deep blank-field catalogue 683 There are some regions in the Southern hemisphere where no DBFs have been found. Some of these DBFs have been validated with the GTC and they are regularly used in the nightly operation of the telescope (see Section 3). Although our main goal in this work was to provide a list of DBFs for data calibration, the DBFs found can also be used for other interesting purposes (e.g. deep pencil-beam extragalactic surveys). In this sense, the galactic extinction in the direction of the sky where the DBFs are located is useful information that would help in the identification of suitable DBFs. Thus, we used the Galactic Dust Reddening and Extinction utility of the NASA/IPAC Infrared Science Archive,7 to estimate the galactic extinction from Schlegel, Finkbeiner & Davis (1998) for each DBF in our catalogue. In this way, five new parameters have been added to the catalogue: E(B V) at the centre of the DBF, the average E(B V) in the sky region defined by the DBF, together with its standard deviation, and the maximum and minimum E(B V) in this sky area. Figure 6. USNO-B Catalog entries superimposed on to the DSS-blue arcmin2 image centred at (α J2000, δ J2000 ) = ( , ). it has included a corresponding BF entry in our catalogue. The same problem occurred with another region of 16 arcmin diameter around (α J2000, δ J2000 ) = (20:12:40, 74:14:39). These two false void-of-stars regions were detected by a change in the visual inspection of the DBFs, so we cannot discard further similar defects in the USNO-B Catalog. Consequently, we cannot rule out the possible presence of spurious BFs in our catalogue. These two detected spurious entries were manually rejected from our catalogue. 2.3 Results The final result is a catalogue that is currently the deepest catalogue of regions that are void of stars. The catalogue is accessible through the TESELA tool (as explained in Section 4). Table 2 presents a quantitative description of the above results. Column 1 gives the threshold magnitude mth and column 2 lists the number of DBF regions found N DBF. Columns 3, 4 and 5 indicate the mean radius ρ, the standard deviation σ ρ and the maximum DBF radius ρ max, respectively. The integrated mean flux Fint, the standard deviation σfint and the minimum integrated flux F min and maximum integrated flux F min are given in columns 6, 7, 8 and 9, respectively. In Fig. 7, we show the maps of the DBFs found in galactic coordinates (IAU 1958). As expected, the major concentration of DBFs is close to the galactic poles, with a low number of DBFs close to the galactic plane. In order to provide a short list of DBFs that can be used at the telescope, we have selected, among the best regions found, one DBF for approximately a half-hour in RA in each hemisphere, as provided in Table 3. The equatorial and galactic coordinates are given first (columns 1 4). The radius ρ of the DBF is given in column 5. The numbers of stars N with mr lower than 17.5 and 22.0 mag are listed in the columns 6 and 7, respectively. Columns 8 and 9 show mr of the faintest and brightest stars in the field. The integrated flux F int is given in column 10. The last four columns give the galactic extinction E(B V) as described below. C 2012 The Authors, MNRAS 427, C 2012 RAS Monthly Notices of the Royal Astronomical Society We have used the 10.4-m GTC, located at the Observatorio Roque de los Muchachos in La Palma (ORM), Canary Islands, Spain, as a test-bed for the validation of our DBFs. The telescope is currently equipped for scientific exploitation with the OSIRIS optical facility, which includes imaging with broad-band filters from the Sloan set (ugriz), as well as a very flexible set of narrow-band filters that can be used when requested. The OSIRIS instrument consists of a mosaic of two Marconi CCD detectors, each with pixels and a total unviggnetted field of view of arcmin2, giving a plate scale of arcsec pix 1. For this reason, the series of BFs requested for the nightly operation of the instrument must be at least 10-arcmin wide, it must be larger than the OSIRIS field of view, and it must also take into account the dithering between consecutive exposures that is usually performed to eliminate the contribution of the stars in producing the final flat-field image. To date, the most comprehensive list of BFs available for observations carried out at the ORM has been the collection made by Azzaro. However, the large aperture of the GTC means that these fields (which are empty up to magnitudes 10 16) are too shallow to be used as BFs for the GTC, because they appear relatively full of stars once an OSIRIS/GTC image is obtained. During the nightly operation of OSIRIS/GTC, an automatic script is used to obtain a series of broad-band flat-field images in Sloan griz filters (the Sloan u filter is rarely used). Because of the particularities of the OSIRIS shutter, in order to obtain a photometric accuracy better than 0.1 per cent in the images, exposure times of the order of 1 2 s are needed. To define the limiting magnitude for the BFs to be used at the GTC, a complete set of the Azzaro fields was observed over nearly a year of operation at the telescope, and they were reduced photometrically to determine the magnitude of the detected stars under good photometric conditions. For 1 2 s of exposure time, in the Sloan r filter, and with nearly ADUs of sky background level (which represents half of the full well of the detector), stars as faint as mag (depending on the seeing conditions) were detected. This is the fainter limit that we have imposed on the BFs to be used with OSIRIS/GTC, which might be empty regions of the sky up to r = 17.5 mag. This is DEEP BLANK FIELDS FOR OSIRIS/GTC

6 684 F. M. Jime nez-esteban et al. Table 2. Results of the tessellation process using the USNO-B Catalog with the R filter. (1) mth (mag) (2) N DBF 11.0a a (3) ρ (arcmin) (4) σρ (arcmin) (5) ρ max (arcmin) (6) Fint (erg s 1 cm 2 A 1 ) (7) σfint (erg s 1 cm 2 A 1 ) (8) F min (erg s 1 cm 2 A 1 ) (9) F max (erg s 1 cm 2 A 1 ) e e e e e e e e e e e e e e e e-15 From the 11 threshold magnitude catalogue of Paper I obtained with the combination of Tycho-2 B and V bands. impossible to achieve with the Azzaro catalogue. In addition, for operating reasons, these BFs should provide complete RA coverage. These two constraints are very restrictive, but these constraints can sometimes be met using our DBF catalogue. Thus, we have been able to create a list of 87 DBFs that can be used with OSIRIS/GTC, although it was not always possible to archive the mth limit. This list provides at least one of these DBFs both at the beginning or at the end of every night of the year, which is extremely useful for the calibration of the scientific data delivered by the OSIRIS instrument. This list has superseded the Azzaro catalogue in the nightly operation of the telescope. Some of the DBFs selected can be found in Table 3. There is an evident advantage to using the new DBFs rather than those from the Azzaro collection, because they have been shown to be much better. In Fig. 8, we show a comparison of one of the DBFs used at the GTC with another from the Azzaro catalogue. the Spanish Virtual Observatory,8 which can be publicly accessed from The tool was implemented with a data base containing the new computed DBF regions from the USNO-B Catalog. Thus, through a user-friendly interface, the user can now perform a cone-search around a position in the sky in either the BF catalogue from Tycho2 (see Paper I for a extended description) or the DBF catalogue from the USNO-B Catalog. Through the search form for the DBF catalogue, users can select mth and the search radius. The conesearch can be made in either equatorial or galactic coordinates. TESELA presents the result of the cone-search in a table containing the same main properties of the DBFs as presented in Table 3. This table can be downloaded in CSV format for further use. TESELA also provides users with the possibility of visualizing the data. To do this, TESELA takes advantage of ALADIN. Thanks to this connection with ALADIN, we have provided the user with the full capacity and power of the VO. Using this ALADIN window, users 4 TESELA In order to provide easy access to the DBF catalogue, we have incorporated it into TESELA, a web-accessible tool, developed within 8 C 2012 The Authors, MNRAS 427, C 2012 RAS Monthly Notices of the Royal Astronomical Society Figure 7. Galactic positions of the DBFs with mth 15, 16, 17 and 18 in the upper-left, upper-right, lower-left and lower-right panels, respectively.

7 Deep blank-field catalogue 685 Table 3. Short list of selected DBFs. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) RA Dec. Gal. long. Gal. lat. ρ N with mr mr Fint E(B V)(mag) (J2000) (J2000) ( ) ( ) (arcmin) <17.5 <22.0 Max Min (erg s 1 cm 2 A 1 ) Centre Mean Max Min Northern hemisphere 00:06: :06: E (72 ± 3)E :42: :20: E (58 ± 10)E :08: :23: E (260 ± 14)E :35: :33: E (49 ± 3)E :01: :06: E (420 ± 5)E :14: :01: E (113 ± 7)E :56: :38: E (184 ± 17)E :33:31.70 a +31:02: E (21 ± 3)E :13: :08: E (169 ± 6)E :26:38.69 a +24:39: E (118 ± 10)E :45: :11: E (81 ± 7)E :31: :37: E (20 ± 3)E :47: :24: E (7 ± 5)E :24: :37: E (70 ± 4)E :12: :17: E (538 ± 6)E :36: :48: E (288 ± 19)E :48:36.72 a +41:31: E (433 ± 9)E :52:36.48 a +24:43: E (324 ± 9)E :08: :08: E (365 ± 10)E :33: :58: E (179 ± 3)E :58:19.68 a +24:59: E (451 ± 4)E :30:25.20 a +33:15: E (163 ± 6)E :00: :03: E (183 ± 5)E :33:39.12 a +13:26: E (424 ± 11)E :54: :00: E (258 ± 6)E :41:44.40 a +21:58: E (293 ± 3)E :56:12.72 a +17:26: E (306 ± 5)E :34: :56: E (120 ± 5)E :02:58.08 a +35:11: E (79 ± 5)E :22: :53: E (129 ± 6)E :55: :30: E (184 ± 4)E :31: :15: E (59 ± 2)E :53: :32: E (155 ± 9)E :30: :17: E (138 ± 7)E :08: :40: E (305 ± 16)E :25: :45: E (269 ± 17)E :57: :56: E (454 ± 6)E :28: :03: E (249 ± 16)E :58: :21: E (134 ± 7)E :31: :20: E (310 ± 11)E :15: :01: E (45 ± 3)E :43: :49: E (71 ± 4)E :59: :58: E (86 ± 5)E :39: :20: E (111 ± 6)E :00:35.28 a +76:30: E (118 ± 2)E

8 686 F. M. Jiménez-Esteban et al. Table 3 continued (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) RA Dec. Gal. long. Gal. lat. ρ N with mr mr Fint E(B V) (mag) (J2000) (J2000) ( ) ( ) (arcmin) <17.5 <22.0 Max Min (erg s 1 cm 2 A 1 ) Centre Mean Max Min 22:21: :07: E (89 ± 3)E :06: :30: E (490 ± 15)E :22: :38: E (413 ± 10)E Southern hemisphere 00:06: :15: E (253 ± 5)E :37: :20: E (198 ± 15)E :59:31.22 a 11:20: E (319 ± 15)E :39: :50: E (249 ± 5)E :09: :31: E (276 ± 8)E :33: :25: E (207 ± 4)E :08: :38: E (221 ± 4)E :36: :58: E (274 ± 8)E :02: :14: E (49 ± 5)E :20: :16: E (56 ± 3)E :01: :54: E (115 ± 4)E :41:55.18 a 08:27: E (351 ± 9)E :57: :37: E (55 ± 5)E :05: :00: E (282 ± 9)E :31: :45: E (336 ± 9)E :10: :15: E (407 ± 10)E :26: :07: E (552 ± 19)E :05: :46: E (191 ± 16)E :23: :42: E (66 ± 3)E :58: :42: E (296 ± 12)E :19: :09: E (234 ± 4)E :55: :47: E (182 ± 9)E :38: :44: E (80 ± 4)E :00: :33: E (47 ± 2)E :37: :14: E (724 ± 3)E :04: :47: E (173 ± 7)E :42: :06: E (861 ± 15)E :53:31.92 a 04:42: E (93 ± 6)E :27: :45: E (23 ± 5)E :50: :22: E (169 ± 11)E :34: :23: E (261 ± 5)E :04: :17: E (8 ± 2)E :28:03.12 a 03:49: E (10 ± 5)E :04: :17: E (96 ± 5)E :32: :26: E (48 ± 4)E :52: :06: E (73 ± 8)E :05: :46: E (107 ± 2)E :22: :17: E (241 ± 6)E :07: :14: E (258 ± 10)E :41:36.24 a 13:28: E (497 ± 17)E :58: :34: E (102 ± 12)E :40:51.12 a 03:53: E (374 ± 12)E a DBF validated and regularly used at the GTC.

9 Deep blank-field catalogue 687 Figure 8. Comparison of two images taken with OSIRIS. These are 2-s exposure images at the Sloan r filter, with a background level of ADUs, of a DBF from our catalogue (left panel) and a BF from the Azzaro catalogue (right panel) in the same region of the sky. The size of both fields is arcmin 2. There is an evident improvement of our DBF compared with the one from the Azzaro catalogue. Note that the two apparent detections, both in the right-upper part in CCD1 (left side of the detector) and in the left-bottom part in CCD2 (right side of the detector), are dust grain effects on the OSIRIS CCDs, and they do not correspond with any star in the field (they are also visible in the Azzaro field). are allowed to load images and catalogues both locally and from the VO, having full access to the whole universe resident in the VO. This can be very helpful to determine the potential influence of relatively bright nebulae and extragalactic sources in the regions sought. TESELA sends to ALADIN the searching area, which is plotted in the first plane with a red circle, and the DBFs, which are shown by blue circles in the second layer. The size of the blue circles corresponds to the size of the DBFs. Finally, ALADIN charges from the VO in a last layer the objects of NGC (Complete New General Catalogue and Index Catalogue of Nebulae and Star Clusters; Sinnott 1997). Note that, for obvious reasons, Solar system objects have been not considered, and they must be taken into account in order to make use of BF regions close to the ecliptic at a given date. ACKNOWLEDGMENTS This work was partially funded by the Spanish MICINN under the Consolider-Ingenio 2010 Program grant CSD : First Science with the GTC. 9 This work was also supported by the Spanish Programa Nacional de Astronomía y Astrofísica under grants AyA and AYA , and by Astro- Madrid 10 under project CAM S2009/ESP This work is based on observations made with the GTC, installed at the ORM. This work has made use of ALADIN developed at the Centre de Donneés Astronomiques de Strasbourg, France. This research has made use of the NASA/IPAC Infrared Science Archive, which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. REFERENCES Barron J. T., Stumm C., Hogg D. W., Lang D., Roweis S., 2008, AJ, 135, 414 Bonnarel F. et al., 2000, A&AS, 143, 33 Cardiel N., Jiménez-Esteban F. M., Alacid J. M., Solano E., Aberasturi M., 2011, MNRAS, 417, 3061 (Paper I) Cepa J. et al., 2000, Proc. SPIE, 4008, 623 Delaunay B., 1934, Sur la sphère vide, Bull. Acad. Sci. USSR, De Muth J. E., 2006, Basic Statistics and Pharmaceutical Statistical Applications, 2nd edn. Chapman & Hall/CRC, Boca Raton, FL, p. 577 Gilliland R. L., 1992, Details of Noise Sources and Reduction Processes, in Howell S. B., ed., ASP Conf. Ser. Vol. 23, Astronomical CCD Observing and Reduction Techniques. Astron. Soc. Pac., San Francisco, p. 68 Høg E. et al., 2000, A&A, 355, L27 Monet D. G. et al., 2003, AJ, 125, 984 Schlegel D. J., Finkbeiner D. P., Davis M., 1998, ApJ, 500, 525 Sinnott R. W., 1997, VizieR Online Data Catalogue, 7118, This paper has been typeset from a TEX/LATEX file prepared by the author.